.jpg)
A quick note before we begin: This article is based on a YouTube transcript. I haven’t independently verified every figure or event. Treat the specific numbers as reported claims, not settled facts. This is not investment advice.
In July 2026, a young investor named Leopold Aschenbrenner was being called the “Warren Buffett of AI.” His hedge fund, Situational Awareness, had returned 439% in just six months. By early July, it managed around $45 billion. By the end of the month, it had shrunk to roughly $10 billion.
The story is not about AI being a bad bet. It’s about what happens when a correct thesis meets reckless leverage.
Who Is Leopold Aschenbrenner?
Aschenbrenner’s background is unusual. He finished high school at 15, graduated as valedictorian from Columbia University at 19, and in 2023 joined OpenAI’s “Superalignment” team—a group working on the safety of superintelligent AI systems.
In 2024, OpenAI fired him. The company said he improperly disclosed internal information. Aschenbrenner has disputed that account, saying he was pushed out after raising internal concerns about security practices. Later that year, he published a 165-page essay called Situational Awareness, outlining how AI could reshape the global economy. The essay went viral in Silicon Valley and made him a known voice on AI’s future.
He then launched a hedge fund with the same name. Early backers reportedly included Stripe co-founders Patrick and John Collison and former GitHub CEO Nat Friedman. The fund’s strategy was concentrated: buy AI infrastructure stocks—chips, memory, data centers, power—and bet against legacy software companies that AI might disrupt.
What Happened?
The fund’s performance was extraordinary. By June 2026, it had returned 439% for the year and over 1,000% since inception. Assets peaked at around $45 billion in the first week of July.
Then the market turned. In late July, AI-related stocks sold off sharply. Sandisk, Micron, SK Hynix, and CoreWeave—all major holdings—fell by more than 35%. Because the fund was running at roughly 4x leverage, a 25% decline in its holdings would wipe out 100% of its equity. The actual losses triggered margin calls from its prime brokers: Bank of America, Citigroup, Goldman Sachs, and JPMorgan Chase.
A margin call means the lender demands more collateral. If the borrower can’t provide it, the lender sells the assets. Aschenbrenner had no new capital coming in—investors were already spooked by the AI selloff. His only option was to sell.
On July 30, Citadel, the hedge fund founded by Ken Griffin, bought the bulk of Situational Awareness’s public equity portfolio—roughly $16 billion worth of stocks—at a discount of more than 10%. The fund’s assets fell from $45 billion to about $10 billion. It retained a stake in [Anthropic](https://www.anthropic.com) valued at around $5 billion.
Why It Happened: The Math of Leverage
The collapse had little to do with AI being a bad investment. The AI infrastructure thesis was largely correct. The problem was the leverage.
Here’s how it works. If you have $1 million and buy $1 million of stock, a 10% gain gives you $100,000. A 10% loss costs you $100,000. But if you borrow another $3 million and buy $4 million of stock—4x leverage—a 10% gain gives you $400,000, a 40% return on your $1 million. A 10% loss costs you $400,000, a 40% loss. If the stock falls 25%, you lose 100% of your money.
This is why Warren Buffett and his late partner Charlie Munger warned against leverage. They had a famous saying about the “three Ls” that can ruin a smart person: liquor, ladies, and leverage. The first two are personal. The third is financial, and it is ruthless.
When the fund’s holdings fell 35% or more, the 4x leverage meant the fund’s equity was wiped out several times over. The prime brokers had no choice but to demand collateral and then liquidate.
The Citadel Rescue—and What It Revealed
Citadel’s purchase was a masterstroke. By buying the positions at a discount, Citadel removed the market’s largest forced seller. Once traders realized the fund’s remaining positions were no longer being liquidated, the AI stocks that had been sold off rebounded sharply. SK Hynix and CoreWeave, two former holdings, rallied after the sale.
Ken Griffin later told investors that Citadel had already sold off about 80% of the risk from the Situational Awareness portfolio. The episode also exposed a structural problem: the fund’s positions were so large and concentrated that the mere knowledge of its distress was enough to move markets. As one strategist put it, the market had been confused for weeks about who was driving a second leg of selling in momentum stocks. The answer, it turned out, was a single leveraged fund.
Key Facts and Figures
Peak assets: Approximately $45 billion in early July 2026.
Six-month return: 439% net return through June 2026.
Since inception return: Over 1,000% from late 2024 through June 2026.
Leverage: Approximately 4x.
Post-collapse assets: About $10 billion.
Citadel purchase: Approximately $16 billion in public equity at a 10%+ discount.
Prime brokers: Bank of America, Citigroup, Goldman Sachs, and JPMorgan Chase.
Retained holding: A stake in Anthropic valued at approximately $5 billion.
The SEC Investigation
In late August 2026, the U.S. Securities and Exchange Commission sent subpoenas to the four major banks that had lent money to Situational Awareness. The SEC is seeking information about the timing of the fund’s trades and its communications with lenders about leverage. The investigation is at an early stage. Neither the fund nor the banks have been accused of wrongdoing. A spokesperson for Situational Awareness said the fund would cooperate fully with any regulatory request.
What This Means for Indian Investors
India’s margin trading facility (MTF) market has grown rapidly. As of January 2026, Zerodha’s outstanding MTF loan book had reached a record ₹1.16 trillion, up nearly 50% year-on-year and more than four times the level four years earlier. Zerodha founder Nithin Kamath has warned that the combination of high leverage, illiquid mid- and small-cap stocks, and weak risk frameworks could lead to “synchronised liquidations” during a sharp market correction.
Kamath has pointed out a crucial structural problem: Indian equities have decent liquidity when markets rise, but liquidity dries up completely during drawdowns. With minimal short-selling, there is almost no natural buyer when prices fall. Forced liquidations become self-reinforcing, especially in stocks that are not part of the futures and options segment.
India’s leverage levels are far lower than the 4x used by Situational Awareness. The Securities and Exchange Board of India caps MTF exposure at 50% of a broker’s net worth. But the fundamental risk remains the same: leverage magnifies gains in good times and destroys capital in bad times.
Important Questions and Uncertainties
Several aspects of the Situational Awareness collapse remain unclear. The full extent of the fund’s hedging strategy—including short positions in software companies like Adobe—has not been publicly detailed. Whether the SEC investigation will lead to any enforcement action is unknown. And it remains to be seen whether other funds in the AI trade are carrying similar levels of hidden leverage.
What is clear is that the AI boom has been increasingly financed by debt. The collapse of Situational Awareness is a reminder that when leverage unwinds, it can happen faster and more violently than almost anyone expects.
Conclusion
Leopold Aschenbrenner was not wrong about AI. He was wrong about leverage. His story is not one of a bad investment thesis—it is a case study in how a correct thesis, combined with excessive borrowing, can still end in disaster. The AI trade may well continue to deliver enormous returns. But the margin call does not care about your thesis. It only cares about collateral.
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg&description=The 24-Year-Old Who Built a $45 Billion AI Hedge Fund—and Lost Most of It in Weeks){kind=link}
0 Comments